The proposed dissertation study will explore how attachment behaviors relate to substance use and examine the effects of treatment on attachment-related behaviors by assessing several facets of attachment-related themes, substance abuse outcome, and type of treatment. Recent work suggests attachment is malleable and changes because of the dramatic developmental psychological and physical changes adolescents undergo (Allen & Land, 1999; Berlin & Cassidy, 1999; Kobak, 1999). Yet little is known about attachment-related behaviors and their role in the development or inhibition of substance-use disorders in adolescents. Moreover, the prevalence of substance abuse is a significant problem in adolescents, and any empirical evidence that can enhance prevention and treatment efforts would be invaluable. The analysis will be designed to measure the impact of attachment behaviors (autonomy and relatedness) and type of treatment (family, individual, and group therapy) on substance-abuse treatment outcome. This study proposes to code 200 videotaped family interactions using the Autonomy and Relatedness Coding System (Allen, Hauser, Bell, McEIhaney et al., 1994). The family interactions will then be used to determine whether attachment mediates the type of treatment and treatment-outcome relationship. The specific aims are: 1) To examine, at baseline, the nature of the relationship between attachment and substance use. 2) To examine the effect of treatment on attachment-related behaviors. 3) To determine whether attachment mediates the type of treatment and treatment-outcome relationship. This study will culminate with an improved understanding of the role of attachment in substance-abuse treatment as well as valuable insights for future research of family attachment as a prevention for adolescent substance abuse. This dissertation project proposes to use data from a completed, NIDA-funded treatment-outcome study, R01 DA09422, and to help illuminate important findings about why certain adolescent substance abuse treatments work.